Artificial Intelligence Assisted Optimization of Ramaria obtusissima Extracts and Their Integrated Chemical and Biological Characterization

Abstract

In this study, the biological activities of extracts obtained from Ramaria obtusissima were optimized using response surface methodology (RSM) and artificial neural networks-genetic algorithm (ANN-GA) approaches, and the chemical and biological profiles of the obtained extracts were evaluated with a holistic approach. Antioxidant potential was determined using FRAP, DPPH, TAS, TOS, and OSI parameters. It was found that the extract optimized with ANN-GA had significantly higher FRAP (242 ± 3 mg Trolox equivalent/g), TAS (6.64 ± 0.04 mmol/L), and DPPH (154 ± 3 mg Trolox equivalent/g) values compared to the RSM extract, while its OSI value was lower. Anticholinesterase activities were evaluated using IC₅₀ values, and it was determined that the ANN-GA extract exhibited a stronger inhibitory effect on acetylcholinesterase (95 ± 2 µg/mL) and butyrylcholinesterase (125 ± 3 µg/mL) compared to the RSM extract. Antiproliferative effects were investigated in A549, MCF-7, and DU-145 cell lines, and a significant and dose-dependent suppression of cell proliferation was observed in all three cell lines, particularly at concentrations of 100 and 200 µg/mL. The chemical profile was determined using LC-MS/MS and GC-MS techniques. Higher levels of phenolic compounds such as gallic acid (6694.5 ± 4.9 mg/kg), caffeic acid (3374.8 ± 4.9 mg/kg), and quercetin (1563.1 ± 2.3 mg/kg) were found in the ANN-GA extract. GC-MS analyses showed that the ANN-GA extract has a richer lipophilic component profile in terms of biologically active fatty acids and ester derivatives. The findings reveal that AI-assisted optimization offers a powerful and effective approach to enhancing the biological efficacy of mushroom-derived natural products.

1. Introduction

Medicinal mushrooms are no longer limited to their traditional uses; today, they are valued as high-value biological resources at the intersection of nutrition science, pharmacology, and biotechnology disciplines. These mushrooms offer a rich nutritional profile with highly bioavailable proteins, complex carbohydrates, digestible dietary fiber, B vitamins, and minerals such as potassium, selenium, and zinc, while their low fat and energy content provides advantages for healthy eating [1,2]. However, the main therapeutic potential of medicinal mushrooms stems from secondary metabolites such as β-glucan polysaccharides, phenolic compounds, flavonoids, triterpenes, and sterols. The immunomodulatory, antioxidant, antimicrobial, and antiproliferative effects of these components have been strongly demonstrated in the literature, and the level of these biological activities largely depends on the conditions applied during the extraction process [3,4,5]. Therefore, extract optimization stands out as a critical step in enhancing biological activity by ensuring the maximum yield and effectiveness of target bioactive compounds. Recent studies have increasingly emphasized that extraction efficiency and the resulting biological activity of mushroom-derived products are strongly influenced by both processing conditions and solvent characteristics. Advanced extraction techniques, including ultrasound-assisted, microwave-assisted, and green solvent-based approaches, have been widely investigated to enhance the recovery of phenolic compounds and other bioactive metabolites from edible and medicinal mushrooms. These studies collectively demonstrate that optimized extraction strategies not only improve antioxidant and enzyme inhibitory activities but also modulate the qualitative chemical profile of fungal extracts, thereby shaping their functional potential [6,7,8].

In this study, optimum extraction conditions that maximize the biological activity of Ramaria obtusissima (Peck) Corner were determined. The antioxidant, anticholinesterase, and antiproliferative activities, as well as the phenolic content, of the extracts obtained under these conditions were evaluated. In addition, the chemical composition profiles and dominant bioactive components of the extracts were characterized in detail. In this context, artificial intelligence-assisted optimization refers to data-driven modeling approaches, including artificial neural networks integrated with evolutionary algorithms, which enable the analysis of nonlinear interactions and multi-parameter optimization in complex biological extraction systems. The findings clearly demonstrate the importance of extract optimization-based approaches for the scientific development of functional food and pharmaceutical products derived from medicinal mushrooms. R. obtusissima is a coralloid macrofungus species with yellowish tones, branched and umbrella-like arms, and is among the taxa that stand out within the Ramaria genus due to its morphological characteristics. Most studies on the species have focused on its ecological and chemical characteristics, such as its taxonomic position, geographical distribution, and heavy metal content, rather than its biological activity potential [9]. Furthermore, distinctive keys have been established for taxa with yellowish fruiting bodies and flat spores within the Laeticolora subgenus of the Ramaria genus [10]. When the distribution and ecological characteristics of the species are examined, it is seen that R. obtusissima is mostly reported as a rare humus saprotroph in different geographical areas. It has been recorded as a very rare species on the soil in pine forests [11]. It has also been reported to grow on the humus layer in mixed oak, beech, and chestnut forests [12]. In this context, it is revealed that R. obtusissima is distributed mostly on the soil and in limited densities in coniferous and mixed forest ecosystems. R. obtusissima is reported to be among the edible wild mushroom species [13].

2. Results and Discussion

2.1. Optimization of Extraction Conditions

In this study, extraction conditions, which play a decisive role in total antioxidant status (TAS), were systematically evaluated. In this context, extraction temperature (30, 45 and 60 °C), processing time (30, 45 and 60 min) and solvent composition (0%, 50% and 100% ethanol/water) were defined as independent variables. The TAS measurement values obtained as a result of extraction applications carried out using different combinations of these parameters were quantitatively determined and the findings are presented in detail in Table 1.

Table 1.

TAS values of the extracts obtained from Ramaria obtusissima.

Experiment Number	Extraction Time (min)	Extraction Temperature (°C)	Ethanol/Water Ratio (%)	TAS (mmol/L)
1	30	30	0	5.172 ± 0.038 j
2	45	30	0	5.655 ± 0.042 m
3	60	30	0	4.209 ± 0.064 b
4	30	30	50	5.426 ± 0.047 l
5	45	30	50	5.969 ± 0.039 o
6	60	30	50	4.533 ± 0.050 c
7	30	30	100	5.496 ± 0.045 lm
8	45	30	100	5.833 ± 0.050 n
9	60	30	100	4.356 ± 0.044 bc
10	30	45	0	5.632 ± 0.047 m
11	45	45	0	5.876 ± 0.036 n
12	60	45	0	4.445 ± 0.039 c
13	30	45	50	5.992 ± 0.038 o
14	45	45	50	6.119 ± 0.040 p
15	60	45	50	4.774 ± 0.045 d
16	30	45	100	5.758 ± 0.036 mn
17	45	45	100	6.037 ± 0.057 o
18	60	45	100	4.781 ± 0.047 d
19	30	60	0	5.628 ± 0.031 m
20	45	60	0	5.822 ± 0.054 n
21	60	60	0	4.256 ± 0.037 b
22	30	60	50	5.674 ± 0.040 m
23	45	60	50	5.892 ± 0.033 n
24	60	60	50	4.366 ± 0.036 bc
25	30	60	100	5.577 ± 0.034 m
26	45	60	100	5.621 ± 0.058 m
27	60	60	100	4.162 ± 0.030 b

Note: Different superscript letters indicate statistically significant differences between groups (p < 0.05). Values sharing the same letter are not significantly different. Values marked with two letters (e.g., ab) indicate overlap between statistical groups and do not differ significantly from groups labeled with either letter.

In this study, when the effects of different extraction times, temperatures, and solvent ratios on TAS values were evaluated together, it was clearly seen that the antioxidant capacity changed in a sensitive manner to the interaction between the parameters. In general, the 45 min extraction time stood out as the time in which the highest TAS values were obtained at all temperatures and solvent ratios. In particular, the extraction performed at 45 °C and a 50% ethanol/water ratio yielded the highest TAS level recorded in the study with a value of 6.119 ± 0.040 mmol/L. This shows that the extraction of antioxidant compounds is more efficient under medium temperature conditions and with intermediate solvent polarity. Similarly, the high TAS value obtained in 45 min with a solvent containing 100% ethanol at the same temperature (6.037 ± 0.057 mmol/L) reveals that the effect of solvent composition becomes more pronounced with time. In contrast, a significant decrease in TAS values was observed in all conditions where the extraction time was extended to 60 min. This decrease became particularly pronounced at 60 °C, with the lowest TAS value (4.162 ± 0.030 mmol/L) recorded in the 60 min–100% ethanol combination. The decrease in the structural stability of antioxidant compounds and the activation of possible degradation processes in combinations of high temperature and long duration can be considered the main reason for this decrease. Indeed, it is observed that the TAS values obtained at 60 °C generally follow a lower range (4.162 ± 0.030–5.892 ± 0.033 mmol/L), regardless of the solvent ratio. The results obtained under low temperature conditions (30 °C) showed a more balanced distribution, but the maximum TAS values lagged behind those under medium temperature conditions. At this temperature, the highest TAS value was determined as 5.969 ± 0.039 mmol/L at 45 min and with 50% ethanol; significant decreases occurred at all solvent ratios when the extraction time was increased to 60 min. Overall, it is understood that reaching extreme values in extraction temperature and time negatively affects TAS, while moderate conditions are more favorable for preserving and maximizing antioxidant capacity.

In this study, two different optimization approaches were applied to interpret the experimental data and determine the optimum extraction conditions. In the modeling process carried out within the scope of Response Surface Methodology (RSM), linear, two-factor interactive (2FI), quadratic, and cubic regression models were evaluated comparatively [9,10]. Statistical analyses revealed that the quadratic model provided the highest fit with the dataset with an R² = 0.985 value. This high coefficient of determination indicates that 98.5% of the variation in total antioxidant status (TAS) values can be explained by independent variables such as extraction temperature, time, and solvent ratio. The results show that the quadratic model strongly and reliably represents the relationships between independent variables and the response variable, and successfully describes the extraction process mathematically. Accordingly, a quadratic polynomial regression model was constructed and evaluated to predict the TAS values of R. obtusissima extracts.

$$\begin{array}{c}TAS=6.18-0.019X_1-0.581X_2+0.052X_3-0.091{X_{1}X}_2-0.082X_1{X}_3\\ -0.001X_2{X}_3-0.287X_1^2-0.857X_2^2-0.175X_3^2\end{array}$$

In the established regression model, extraction temperature (X₁), extraction time (X₂), and ethanol/water ratio (X₃) were defined as coded independent variables. To examine the individual and interactive effects of these parameters on TAS in more detail, three-dimensional response surface plots were created and are shown in Figure 1. These plots show that the response variable increases significantly, especially at moderate temperatures and solvent ratios, and that the interactions between the parameters play a significant role in TAS.

Figure 1.

Response surface plots.

In the second phase of the study, an Artificial Neural Network (ANN) algorithm was used to model the experimental results based on predictions. As a result of comprehensive performance evaluations conducted on different network topologies, the most successful structure in terms of prediction accuracy was selected, and this structure was integrated with a Genetic Algorithm (GA) to develop a hybrid optimization model [14,15]. In this context, the ANN model with a 3-4-1 architecture containing four neurons in a single hidden layer was determined to be the most suitable structure. The performance of the selected model was validated with MSE = 0.001, MAPE = 0.399%, and correlation coefficient R = 0.998, indicating that the ANN model has high prediction accuracy and strong generalization ability. In the GA-based optimization process, a search was performed in the solution space using the prediction outputs generated by the ANN model as a reference. Population size, one of the key parameters affecting the algorithm’s performance, was tested at different values, and as a result of the comparisons, it was determined that the most suitable number of individuals was 12. The convergence graph presented in Figure 2 shows that the algorithm reaches a stable solution after approximately 10 iterations, and the optimization process is effectively completed. These findings demonstrate that the ANN-GA-based hybrid approach is a powerful tool providing high accuracy and computational efficiency in optimizing extraction processes.

Figure 2.

Convergence Graph.

2.2. Antioxidant Activity

Natural mushrooms are considered biological materials that make significant contributions to antioxidant defense mechanisms through their phenolic content, structural polysaccharides, and various secondary metabolites. These components enhance the functional and biological value of mushrooms by effectively suppressing reactive oxygen species and limiting oxidative damage [16,17]. In this context, the antioxidant profiles of optimized extracts obtained from R. obtusissima were experimentally investigated. The analysis results are summarized in Table 2.

Table 2.

Antioxidant parameters of Ramaria obtusissima optimized extract.

Parameters	ANN-GA Extract Values	RSM Extract Values
FRAP (mg Trolox Equi/g)	242.43 ± 2.75 a	207.21 ± 3.83 b
DPPH (mg Trolox Equi/g)	153.98 ± 2.72 a	129.78 ± 2.14 b
TAS (mmol/L)	6.638 ± 0.041 a	6.219 ± 0.044 b
TOS (µmol/L)	10.134 ± 0.081 b	11.142 ± 0.081 a
OSI (TOS/(TAS × 10))	0.153 ± 0.002 b	0.179 ± 0.003 a

Note: Means with different superscript letters in the same column are significantly different according to Duncan’s multiple range test (p < 0.05).

In this study, the antioxidant profiles of optimized extracts obtained from R. obtusissima were evaluated in detail, and significant differences were revealed depending on the optimization approach used. The statistically higher FRAP, TAS, and DPPH values of the ANN-GA-optimized extract compared to the RSM extract indicate that this approach creates a more advantageous extraction profile in terms of reducing power, total antioxidant capacity, and free radical scavenging efficiency. In contrast, the higher TOS and consequently OSI values in the RSM extracts reveal that the oxidant load is more dominant in these extracts and that the antioxidant–oxidant balance exhibits a relatively more disadvantageous structure. In particular, the lower OSI value calculated in the ANN-GA extract points to a redox balance associated not only with increased antioxidant capacity but also with more effective suppression of oxidative stress potential. The absence of any findings in the literature directly related to the antioxidant activity of R. obtusissima necessitates an evaluation of the results obtained in this study in relation to other species of the Ramaria genus. Previous studies have reported significant antioxidant activities in different Ramaria species such as Ramaria flava, Ramaria stricta, Ramaria subalpina, and Ramaria patagonica, indicating that this genus generally has a biochemical profile rich in redox-active compounds [18,19,20,21,22]. In this context, the high FRAP, TAS, and DPPH values observed in R. obtusissima extracts exhibit a profile consistent with the general antioxidant trend reported in Ramaria species.

No data regarding the TAS, TOS, and OSI values of R. obtusissima were found in the literature. In this context, TAS (4.223 mmol/L), TOS (8.201 µmol/L), and OSI (0.194) values reported for Ramaria stricta in studies on different Ramaria species have been reported [21]. The fact that the TAS values of the optimized extracts obtained from R. obtusissima used in our study were found to be higher than those of R. stricta in both optimization approaches shows that this species has a remarkable potential in terms of total antioxidant capacity. In particular, the lower OSI value calculated in the ANN-GA extract reveals that the increased antioxidant capacity suppresses the oxidant load more effectively and offers a more biologically balanced redox profile. Furthermore, studies on different wild mushrooms have reported that Paralepista flaccida has a TAS value of 4.054 mmol/L, a TOS value of 10.352 µmol/L, and an OSI value of 0.255 [14]. In Phylloporia ribis, the TAS value was reported as 6.092 mmol/L, the TOS value as 13.389 µmol/L, and the OSI value as 0.220 [15]. Russula grata has been reported to have a TAS value of 3.718 mmol/L, a TOS value of 10.410 µmol/L, and an OSI value of 0.280 [23]. In Hericium erinaceus, the TAS value was determined as 5.426 mmol/L, the TOS value as 6.621 µmol/L, and the OSI value as 0.122 [24]. Compared to these studies, the optimized extracts obtained from Ramaria obtusissima used in our study reached higher values than many wild mushroom species, especially in terms of TAS values, while the OSI values were lower or comparable. In particular, the high TAS and low OSI profile exhibited by the ANN-GA-optimized extract shows that the antioxidant capacity suppresses the oxidant load more effectively and that R. obtusissima has an advantageous biochemical profile in terms of antioxidant–oxidant balance. Considering that the TAS value is a general indicator of antioxidant compounds synthesized within the mushroom, and the TOS value reflects the total of oxidant compounds [25], the low OSI values observed in R. obtusissima extracts, especially after ANN-GA optimization, indicate that oxidant compounds are highly suppressed by endogenous antioxidants. In this context, it reveals that R. obtusissima possesses a strong endogenous antioxidant defense system and that the optimization strategy used significantly shapes this biochemical balance. In conclusion, the antioxidant potential of R. obtusissima shows that ANN-GA-based optimization can enable the achievement of more balanced and biologically advantageous antioxidant profiles in Ramaria species.

2.3. Anticholinesterase Activity

Mushrooms, which are rich in bioactive components, are considered among the natural resources that can affect the cholinergic system through the inhibition of acetylcholinesterase and butyrylcholinesterase enzymes. This inhibitory effect increases the pharmacological potential of mushrooms in terms of protecting nerve transmission and supporting biochemical mechanisms associated with neurodegenerative processes [26,27]. In this context, the anticholinesterase activities of optimized extracts obtained from R. obtusissima were experimentally investigated, and the calculated IC₅₀ values are presented in Table 3.

Table 3.

Anticholinesterase activity of Ramaria obtusissima optimized extract.

Sample	AChE μg/mL	BChE μg/mL
Galantamine	7.58 ± 0.34 c	17.30 ± 0.36 c
ANN-GA extract	94.81 ± 2.21 b	125.11 ± 3.13 b
RSM extract	104.98 ± 2.28 a	157.70 ± 3.06 a

Note: Means with different superscript letters in the same column are significantly different according to Duncan’s multiple range test (p < 0.05).

In our study, the anticholinesterase activities of optimized extracts obtained from R. obtusissima were evaluated based on IC₅₀ values. While both optimized extracts showed measurable activity in terms of AChE and BChE inhibition, the ANN-GA-optimized extract exhibited lower IC₅₀ values for both enzymes compared to the RSM extract, indicating a stronger inhibitory effect. Specifically, the lower IC₅₀ value of the ANN-GA extract for AChE inhibition suggests that this optimization approach provides a more effective extraction profile in terms of cholinergic system-related biological activity. The higher IC₅₀ values obtained in the RSM extracts, however, suggest that the extraction conditions may have limited the effectiveness of the inhibitory components. While it is expected that galantamine, used as a reference inhibitor, would exhibit significantly lower IC₅₀ values for both enzymes, it is noteworthy that R. obtusissima extracts, as naturally occurring samples, showed significantly higher levels of anticholinesterase activity. The combined evaluation of AChE and BChE inhibition reveals that the ANN-GA extract offers a more balanced and effective inhibition profile on both enzymes. This suggests that bioactive compounds capable of interacting with cholinergic enzymes may have been extracted more effectively under optimized extraction conditions. Many naturally occurring mushrooms have been reported in the literature to exhibit acetylcholinesterase and butyrylcholinesterase inhibitory activities, highlighting the biological potential of mushrooms in relation to the cholinergic system [28,29,30]. In line with this general trend, the inhibitory effect of the optimized extracts obtained from R. obtusissima on both enzymes in the present study indicates that this species can be evaluated among fungal-derived anticholinesterase agents. Although the IC₅₀ values of the extracts were higher compared to galantamine, considering the multiple interaction mechanisms of complex natural compounds, the obtained anticholinesterase activities are considered to be biologically significant. Overall, the results reveal that R. obtusissima is a potential natural resource in terms of biological activities related to the cholinergic system, and that ANN-GA-based optimization, in particular, offers a more advantageous approach in enhancing this activity.

2.4. Antiproliferative Activity

Mushrooms are considered natural biological materials that can limit cellular proliferation thanks to the phenolic compounds and various secondary metabolites they contain. These bioactive components can exert an antiproliferative effect through the control of the cell cycle and the regulation of molecular pathways associated with apoptosis [31,32]. Accordingly, the biological effects of optimized extracts obtained from R. obtusissima species on A549, MCF-7 and DU-145 cell lines were investigated under in vitro conditions, and the results obtained are shown in Figure 3.

Figure 3.

Antiproliferative activity of Ramaria obtusissima optimized extract. (In cell culture experiments, control group cells were incubated only in standard culture medium and were not treated with any chemical agent. In the DMSO group, DMSO was added to the cell medium to provide solvent control, without the application of the extract. In the treatment groups, cells were treated with extract solutions prepared at concentrations of 25, 50, 100, and 200 µg/mL, and dose-dependent effects were evaluated. Error bars represent mean ± SD (n = 3)).

In our study, the antiproliferative effects of optimized extracts obtained from R. obtusissima on A549, MCF-7, and DU-145 cell lines were evaluated under in vitro conditions. The observation of similar and high OD values in all cell lines in the control and DMSO groups demonstrates that the solvent used does not have a significant toxic effect on cell viability and reveals the reliability of the experimental system. In contrast, a significant decrease in OD values occurred in the extract-treated groups in parallel with increasing concentration, indicating that R. obtusissima extracts suppress cell proliferation in a dose-dependent manner. The significant decrease in OD observed in all three cell lines, particularly at concentrations of 100 and 200 µg/mL, suggests that the extracts exhibit a stronger antiproliferative effect at higher concentrations. Furthermore, it is noteworthy that the extract obtained with ANN-GA produced lower OD values compared to the RSM extract at the same concentrations. This finding suggests that ANN-GA-based optimization offers a more effective approach in the extraction of bioactive components that suppress cell proliferation. While a generally similar dose-dependent response trend was observed among cell lines, the differences in the level of reduction in OD values depending on the cell type indicate that the extracts may elicit cell-specific biological responses. This may be due to differences in the proliferation rate, metabolic activity, and stress response mechanisms of different cancer cell lines.

The literature reports that species belonging to the genus Ramaria exhibit significant antiproliferative effects against various cancer cell lines. It has been reported that methanol and ethyl acetate extracts obtained from Ramaria botrytis inhibit cell growth in the human colon adenocarcinoma cell line HT-29 and the human hepatocarcinoma cell line HepG2, and that this effect occurs without causing significant cytotoxicity in normal human liver cells [33]. Similarly, it has been reported that Ramaria flava ethanol extracts exhibit a high antiproliferative effect on the human breast cancer cell line MDA-MB-231, with an IC₅₀ value of 66.54 µg/mL, while showing moderate antiproliferative activity in NCI-H520 (lung) and BGC-803 (stomach) cancer cell lines [19]. These literature findings reveal that Ramaria species have significant potential in terms of bioactive compounds that can suppress cancer cell proliferation. In this context, the dose-dependent antiproliferative effects observed in R. obtusissima extracts in our study show a trend consistent with previously reported findings for the Ramaria genus. In particular, the fact that the extract optimized with ANN-GA showed a more pronounced proliferation-suppressing effect in different cancer cell lines indicates that the optimization strategy significantly affects not only the extraction yield but also the cellular biological activity. Overall, the results indicate that R. obtusissima is a species that supports the antiproliferative potential of the Ramaria genus and can exhibit strong and dose-dependent antiproliferative activities in different cancer cell lines with appropriate optimization approaches.

2.5. LC-MS/MS-Based Phenolic Composition

Phenolic compounds are considered among the main phytochemical groups that play a significant role in shaping fungal-derived biological activities. Thanks to their electron-donating and binding capacities, these compounds suppress free radical formation, can form a complex with metal ions, and limit the progression of oxidative reactions [34,35]. In this context, the phenolic component profiles of optimized extracts obtained from R. obtusissima were analyzed using LC-MS/MS technique, and the findings are presented in Table 4.

Table 4.

Phenolic contents of Ramaria obtusissima optimized extract.

Phenolic Compounds	LOD	LOQ	RSM Extract (mg/kg Extract)	ANN-GA Extract (mg/kg Extract)
Gallic acid	2.17	7.24	5172.53 ± 6.91 b	6694.49 ± 4.87 a
Protocatechuic acid	2.21	7.37	1628.58 ± 2.49 b	1823.30 ± 2.84 a
Caffeic acid	2.19	7.29	2144.03 ± 3.94 b	3374.78 ± 4.86 a
2-Hydroxycinnamic acid	1.37	4.56	588.57 ± 3.69 b	808.57 ± 2.09 a
Quercetin	4.18	13.93	1258.81 ± 5.90 b	1563.07 ± 2.33 a
Vanillic acid	24.74	82.46	923.82 ± 2.40 b	1106.81 ± 3.52 a

Note: Means with different superscript letters in the same column are significantly different according to Duncan’s multiple range test (p < 0.05). LOD: limit of detection; LOQ: limit of quantification. LOD and LOQ values are expressed in mg/kg.

In our study, the phenolic component profiles of optimized extracts obtained from R. obtusissima were analyzed in detail using LC-MS/MS techniques. The analyses revealed the presence of different phenolic compounds in the extracts obtained with both optimization approaches, but all identified phenolic compounds were found in statistically higher concentrations in the extracts optimized with ANN-GA. This suggests that the ANN-GA approach is a more effective method than RSM in increasing the extraction efficiency of phenolic compounds. It is noteworthy that among the phenolic compounds identified in R. obtusissima, gallic acid, caffeic acid, and quercetin were found in significantly higher amounts in the ANN-GA extract. It is known that hydroxyl group-rich phenolic acids such as gallic acid and caffeic acid contribute significantly to the antioxidant activity of extracts due to their strong reducing capacities and free radical scavenging properties [36,37]. Similarly, quercetin, a flavonoid, has been reported to play a critical role in biological activities with both its antioxidant and enzyme inhibitor properties [38]. In this context, the high phenolic content observed in the ANN-GA extract provides a biochemical basis consistent with the high FRAP, TAS, and DPPH values determined in our study. The detection of other phenolic compounds such as protocatechuic acid, vanillic acid, and 2-hydroxycinnamic acid at higher levels in the ANN-GA extract indicates that this extract exhibits a multi-component and balanced phenolic profile. It is thought that these compounds may strengthen antioxidant defense through synergistic effects and contribute to the suppression of oxidative stress [39,40,41]. Indeed, the lower TOS and OSI values calculated in the ANN-GA extract support the idea that the increased phenolic content more effectively balances the oxidant load.

Given that phenolic compounds also have regulatory effects on cell proliferation and enzyme activity [37,38], it is considered that the high phenolic concentrations detected in the ANN-GA extract may be related to anticholinesterase and antiproliferative activities. Especially considering the biological effects of compounds such as quercetin and caffeic acid associated with cell cycle regulation, activation of apoptotic signals, and inhibition of cholinergic enzymes [37,38], it can be said that the chemical basis of the strong biological activities observed in R. obtusissima extracts is largely due to phenolic components. In conclusion, LC-MS/MS analyses reveal that ANN-GA-based optimization is more effective in enriching phenolic compounds in R. obtusissima extracts, and this is in complete agreement with antioxidant, anticholinesterase, and antiproliferative results. The findings demonstrate that R. obtusissima is a rich natural resource in terms of phenolic compounds and that its biological activity can be significantly enhanced with appropriate optimization strategies.

2.6. GC-MS Based Profile of Volatile Compounds in Extracts

The biological efficacy of mushroom-derived extracts is closely related to the diversity and distribution of their chemical components. These components may include fatty acids, aromatic compounds, aliphatic hydrocarbons, esters, and various groups of secondary metabolites. These compounds play a fundamental role in the manifestation of the extracts’ biological properties, such as their antioxidant defense capacity, inhibitory effects on cholinesterase enzymes, and suppression of cellular proliferation [42,43]. Accordingly, the chemical profiles of optimized extracts obtained from R. obtusissima were characterized by GC-MS analysis, and the results for the identified components are presented in Table 5. The differences in chemical composition between RSM and ANN-GA-optimized extracts are also clearly reflected in their GC-MS total ion chromatograms (Figure 4). Compounds tentatively identified based on GC–MS spectral matching.

Table 5.

Chemical composition of extracts obtained from Ramaria obtusissima.

Compound Name	Chemical Class	R. Time (min)	Area (%)–RSM	Area (%)–ANN-GA
Caryophyllene	Sesquiterpene	11.738	2.20	2.28
Benzaldehyde, 4-(1-methylethyl)-	Aromatic aldehyde	14.304	1.42	–
Anethole	Phenylpropanoid	15.129	27.24	–
Caryophyllene oxide	Oxygenated sesquiterpene	18.496	2.47	–
2-Pentadecanone, 6,10,14-trimethyl-	Ketone	22.190	3.58	–
Phenol, 2-methoxy-3-(2-propenyl)-	Phenolic compound	22.796	3.07	–
Eicosane	Alkane	27.296	1.51	6.13
Hexatriacontane	Alkane	32.762	6.77	3.90
Butanoic acid, 2-methyl-, 2-methoxy-4-(2-propenyl)phenyl ester	Ester	33.006	2.06	–
2-Hexadecen-1-ol, 3,7,11,15-tetramethyl-	Terpenoid alcohol	35.259	3.31	–
Tetradecanoic acid	Fatty acid	37.271	2.24	9.12
n-Hexadecanoic acid	Fatty acid	42.570	41.53	40.06
Benzene, 1-methoxy-4-(1-propenyl)-	Aromatic compound	–	–	3.57
9,12-Octadecadienoic acid (Z,Z)-, methyl ester	Fatty acid ester	–	–	2.96
Tetracosane	Alkane	–	–	9.26
Ethyl linoleate	Fatty acid ester	–	–	3.03
Docosane	Alkane	–	–	1.78
Oxacycloheptadec-8-en-2-one (8Z)	Lactone	–	–	2.47
9,12-Octadecadienoic acid (Z,Z)-	Polyunsaturated fatty acid	–	–	10.91

Figure 4.

Comparative GC–MS TIC chromatograms of RSM- and ANN-GA-optimized Ramaria obtusissima extracts. (Identifications are tentative and based on spectral library matching. In the chromatograms, the x-axis represents retention time (min), the y-axis represents relative ion intensity (%), and each peak corresponds to an individual volatile or semi-volatile compound detected in the Ramaria obtusissima extracts. ANN-GA: 1, Caryophyllene; 2, Benzene, 1-methoxy-4-(1-propenyl); 3 and 4, Eicosane; 5, 9,12-Octadecadienoic acid (Z,Z)-, methyl ester; 6 and 13, Tetracosane; 7, Ethyl linoleate; 8, Docosane; 9, Oxacycloheptadec-8-en-2-one; 10, Tetradecanoic acid; 11 Hexatriacontane; 12, n-Hexadecanoic acid; 14, 9,12-Octadecadienoic acid (Z,Z). RSM: 1, Caryophyllene; 2, Benzaldehyde, 4-(1-methylethyl); 3, Anethole; 4, Caryophyllene oxide; 5, 2-Pentadecanone, 6,10,14-trimethyl-; 6, Phenol, 2-methoxy-3-(2-propenyl); 7, Eicosane; 8, Hexatriacontane; 9, Butanoic acid, 2-methyl-, 2-methoxy-4-(2-propenyl)phenyl ester; 10, 2-Hexadecen-1-ol, 3,7,11,15-tetramethyl-; 11, Tetradecanoic acid; 12, Hexatriacontane; 13, n-Hexadecanoic acid).

In our study, the chemical profiles of RSM and ANN-GA-optimized extracts obtained from R. obtusissima were characterized by GC-MS analysis. The analysis results reveal that both extracts have a multi-component chemical structure, but the diversity of components and their distribution ratios showed significant differences depending on the optimization strategy used. N-hexadecanoic acid (palmitic acid), identified as the dominant component in both extracts, had the highest percentage of total area and stands out as one of the key components reflecting the lipophilic character of the extracts. Previous studies have reported that palmitic acid can exhibit antioxidant and cellular metabolism-related effects in biological systems [44], suggesting that this component may contribute to the biological activities of the extracts. The detection of anethole, caryophyllene, caryophyllene oxide, and various phenolic and terpenoid compounds in the RSM extract indicates that this extract exhibits a richer profile in terms of aromatic and terpenic components. Considering that compounds such as anethole and caryophyllene are structures associated with antioxidant and biological activity [45,46], it is assessed that certain biological effects of the RSM extract may be related to these components. However, the absence of these components in the ANN-GA extract suggests that optimization approaches create different selectivities in the extraction of volatile and semi-volatile compounds. In the chemical profile of the extract optimized with ANN-GA, the higher concentrations of tetradecanoic acid, 9,12-octadecadienoic acid (linoleic acid), and their ester derivatives, along with various long-chain alkanes, are noteworthy. It is known that polyunsaturated fatty acid derivatives, particularly linoleic acid and ethyl linoleate, can exhibit regulatory effects on antioxidant activity and cell proliferation [47]. The presence of these compounds in higher proportions in the ANN-GA extract provides a chemical basis consistent with the higher antioxidant capacity, lower OSI values, and stronger antiproliferative activity reported in previous sections. Furthermore, the presence of aromatic compounds such as benzene and 1-methoxy-4-(1-propenyl), and lactone structures such as oxacycloheptadec-8-en-2-one detected in the ANN-GA extract, indicates that this extract exhibits a wider chemical diversity. Considering that such compounds may play indirect regulatory roles on cellular stress responses and enzyme activity [48], it is assessed that the anticholinesterase and antiproliferative activities observed in the ANN-GA extract arise through a multi-component and synergistic mechanism of action. When considered together with the high phenolic content and biological activity results determined by GC-MS data and LC-MS/MS analyses, ANN-GA-based optimization demonstrates that it enriches not only the phenolic components but also the biologically active lipophilic components. This holistic chemical profile explains why the ANN-GA extract exhibits stronger biological performance in terms of antioxidant capacity, cholinergic enzyme inhibition, and cell proliferation suppression effects. In conclusion, it reveals that the biological activities of R. obtusissima extracts originate from a multi-component chemical matrix and that the optimization strategy significantly shapes the distribution of these components.

2.7. Contribution of Optimization-Driven Phenolic Profiles to Biological Activities

A key strength of the present study lies in the artificial intelligence-assisted optimization of the extraction process, which substantially influenced the phenolic composition and, consequently, the biological activities of R. obtusissima extracts. Rather than attributing the observed bioactivities to a single compound, the results indicate that optimized extraction conditions selectively enriched specific phenolic groups that collectively determine antioxidant, anticholinesterase, and antiproliferative effects. In particular, optimization resulted in increased levels of low-molecular-weight phenolic acids, including gallic acid, protocatechuic acid, caffeic acid, vanillic acid, and 2-hydroxycinnamic acid, which are well recognized for their strong antioxidant and redox-modulating properties [49,50]. Accordingly, the significantly enhanced DPPH, FRAP, and TAS values observed in the optimized extracts can be interpreted as a direct outcome of optimization-driven modulation of the phenolic acid profile, promoting synergistic radical scavenging mechanisms [51,52]. In parallel, optimization favored the enrichment of flavonoids such as quercetin, luteolin, myricetin, and kaempferol, which are widely associated with cholinesterase inhibition and antiproliferative activity [53,54]. Overall, these findings demonstrate that biological activity in R. obtusissima is primarily governed by optimization-driven shifts in phenolic composition rather than by the presence of individual dominant molecules.

The strengthened radical scavenging and enzyme inhibitory activities observed in the ANN-GA-optimized extract are consistent with the increased levels of key phenolic compounds, particularly gallic acid, caffeic acid, and quercetin, whose hydrogen- and electron-donating capacities directly contribute to free radical neutralization and total antioxidant status [36,37,38]. In addition, their polyhydroxylated structures enable effective interactions with the active sites of acetylcholinesterase and butyrylcholinesterase, supporting the enhanced enzyme inhibitory effects observed [55]. The superior performance of ANN-GA compared to RSM can be attributed to its ability to capture nonlinear interactions between extraction temperature and solvent composition, where slight increases in temperature and ethanol ratio enhance mass transfer and adjust solvent polarity, favoring the simultaneous enrichment of multiple bioactive phenolics. As a result, ANN-GA optimization promotes a phenolic profile that is more favorable for antioxidant, anticholinesterase, and antiproliferative activities. While the anticholinesterase IC₅₀ values of the optimized extracts are noteworthy, they remain substantially higher than that of the reference drug galantamine; therefore, these extracts should be regarded as potential functional ingredients or natural lead candidates rather than direct therapeutic alternatives.

3. Materials and Methods

The macrofungus specimens examined in this study were collected during systematic field studies conducted within the borders of Ankara province, Turkey, in 2025. The collected specimens were assigned the collector number MS-370. Following collection, the samples were transported under appropriate conditions to ensure accurate morphological and taxonomic evaluations and were subsequently deposited in the fungarium collection of Ankara University. Species identification was carried out based on detailed macroscopic and microscopic morphological characteristics by Dr. Mustafa Sevindik and Dr. Ilgaz Akata, both experienced mycologists. The specimens were archived in accordance with the standard registration, labeling, and preservation protocols of the relevant fungarium and preserved for future reference and further analyses.

3.1. Extraction Procedures

Prior to extraction, fruiting bodies of R. obtusissima were dried and ground into a fine powder. This powdered material was used for all extraction experiments conducted within the optimization design. In this study, the experimental process was planned using a multivariate optimization approach to determine the conditions that would maximize extraction efficiency. In this context, three main variables-extraction temperature, application time, and solvent composition-were considered, and each variable was defined at three different levels. The experimental design was based on a full factorial design, allowing for the evaluation of inter-factor interactions, and a total of 27 different experimental combinations were applied. Extraction processes were carried out using an ultrasonic-assisted system. The experiment was configured with a temperature of 30–60 °C, a duration of 30–60 min, and an ethanol concentration ranging from 0% to 100%. The data obtained from the experiments were first modeled using Response Surface Methodology, and the optimum extraction conditions were statistically determined. In addition, Artificial Neural Network and Genetic Algorithm-based approaches were integrated into the optimization process to more effectively explain nonlinear relationships. Thus, a holistic optimization model with high predictive capacity, capable of evaluating the simultaneous effects of multiple variables, was developed.

3.2. Response Surface Methodology (RSM)

In this study, Response Surface Methodology (RSM) was used to ensure that the extraction process was carried out under optimum conditions [14,15]. In the experimental design, extraction temperature, application time, and ethanol/water ratio were determined as independent variables; the effects of these variables on the Total Antioxidant Status (TAS) values of the extracts were evaluated through the response variable. The optimization and modeling process was carried out using Design-Expert software (version 13, Stat-Ease Inc., Minneapolis, MN, USA), which allows for the statistical analysis of experimental data. The obtained data were analyzed using a second-order polynomial regression model to reveal linear and nonlinear relationships between the variables and the response.

$$Y_k={\beta }_{k0}+\sum _{i=1}^n{\beta }_{ki}{x}_i+\sum _{i=1}^n{\beta }_{kii}{x}_i^2+\sum _{i=1}^{n-1}\sum _{j=i+1}^n{\beta }_{kij}{x}_i{x}_j$$

In the established mathematical model, the dependent variable (Y_k) represents the TAS values of the extracts, while the coded independent variables (X_i) represent the extraction temperature, processing time, and solvent composition. The constant coefficient (β_k0) in the model is defined as the parameter reflecting the central point of the experimental design. The statistical validity and fit of the model were evaluated in detail using key indicators such as the coefficient of determination (R²), analysis of variance (ANOVA), and significance levels (p-value). To determine the optimum extraction conditions, critical points were calculated by taking the mathematical derivatives of the model equations, and regions where the response function was maximized were identified. In addition, three-dimensional response surface graphs were created to more clearly reveal the pairwise interactions between the independent variables, and the results obtained were interpreted scientifically through these graphs.

3.3. Artificial Neural Network-Genetic Algorithm (ANN-GA)

All artificial neural network (ANN) modeling procedures were implemented using MATLAB R2023a (MathWorks Inc., Natick, MA, USA), including network training, validation, testing, and performance evaluation. The genetic algorithm (GA) optimization was also performed using the Global Optimization Toolbox of MATLAB. In this study, an artificial intelligence-based predictive modeling approach was adopted to predict the response variable of the extraction process, and an Artificial Neural Network (ANN) was used [14,15]. In the model structure, extraction temperature, application time, and ethanol/water ratio were defined as independent input parameters, while the Total Antioxidant Status (TAS) value was considered as the model output. To increase the generalization ability of the model and reduce the risk of overfitting, the dataset was randomly selected and divided into three subgroups; 80% of the data was used in the training phase, 10% in the validation phase, and 10% in the testing phase. The Levenberg–Marquardt (LM) algorithm was preferred during the training of the network due to its fast convergence feature and success in error minimization. To determine the optimum structure of the ANN architecture, different network topologies were created by varying the number of neurons in the hidden layer between 1 and 20 and were systematically tested. For all network structures, the learning coefficient and momentum value were kept constant, and the maximum number of iterations was set to 500. In the validation process, the early stopping criterion was defined as 50 iterations, and the error tolerance was set at 1 × 10⁻⁵. Each network architecture was evaluated within 1000 independent training cycles with random initial weights and compared based on performance metrics. To quantitatively demonstrate model performance, Mean Squared Error (MSE) and Mean Absolute Percentage Error (MAPE) were used as the basic error criteria. These metrics were calculated according to the mathematical expressions given below:

$$\mathrm{M}\mathrm{S}\mathrm{E}={\displaystyle \frac{1}{\mathrm{n}}}\sum _{\mathrm{i}=1}^{\mathrm{n}}{\left({\mathrm{e}}_{\mathrm{i}}-{\mathrm{p}}_{\mathrm{i}}\right)}^2$$ (1)

$$\mathrm{M}\mathrm{A}\mathrm{P}\mathrm{E}={\displaystyle \frac{1}{\mathrm{n}}}\sum \left|{\displaystyle \frac{{\mathrm{e}}_{\mathrm{i}}-{\mathrm{p}}_{\mathrm{i}}}{{\mathrm{e}}_{\mathrm{i}}}}\right|\times 100$$ (2)

Here, e_i represents the experimentally obtained actual TAS values, and pi represents their estimated equivalents by the ANN model. The total number of observations is denoted by the parameter nnn. The Genetic Algorithm (GA) was used as a computational optimization tool to determine optimum extraction conditions. In the GA application, the effects of the algorithm on solution quality were comparatively examined by considering different population sizes. The selection of new individuals was performed using the probabilistic-based roulette wheel selection method, and a single-point crossover strategy was applied to preserve genetic diversity. The convergence behavior of the algorithm was monitored from the convergence plot graphs obtained during the iterations, and the effect of the number of iterations on solution success was evaluated. To increase the probability of reaching the optimum solution and to reinforce the reliability of the results, each optimization process was repeated at least thirty times.

3.4. Extraction for Biological Activities

Extraction procedures described in this section were performed using the optimized conditions obtained from the optimization stage (Section 3.1) and were specifically designed to produce extracts for biological activity assays. In this study, the extraction process was evaluated using multivariate optimization approaches to maximize the biological activity of extracts obtained from mushroom samples. Optimization analyses performed using RSM resulted in ideal extraction conditions of 42.063 °C temperature, 44.101 min processing time, and 35.084% ethanol/water ratio. In contrast, an Artificial Neural ANN-GA based hybrid model predicted optimum parameters of 44.7990 °C temperature, 38.2150 min extraction time, and 48.6739% ethanol/water ratio. These theoretical optimum values obtained from both models were evaluated together, and in experimental applications, extraction processes were carried out based on parameters closest to the calculated values and applicable under laboratory conditions. For this purpose, an ultrasonically assisted extraction system was used. The processes were carried out at a frequency of 40 kHz, 100% power, and 400 W capacity to ensure methodological consistency in the experimental process. Ultrasonication frequency and power were intentionally kept constant to isolate the effects of extraction time, temperature, and solvent composition, which were selected as the primary optimization variables based on preliminary trials and literature evidence. Following extraction, solvents were removed under reduced pressure using a rotary evaporator at 40 °C until constant weight was achieved. The resulting solvent-free extracts were subsequently used for all biological activity assays.

3.5. Antioxidant Activity Tests

The total antioxidant capacities of optimized mushroom extracts were determined using a commercial analysis kit via the Total Antioxidant Status (TAS) parameter. This analysis is based on the principle of reduction of the 2,2′-azinobis(3-ethylbenzothiazoline-6-sulfonate) (ABTS) radical cation by antioxidant compounds present in the extracts. The decrease in color intensity resulting from the reaction was monitored spectrophotometrically, and the antioxidant capacity was expressed in mmol Trolox equivalent/L [56]. In addition, the total oxidant load of the extracts was evaluated by Total Oxidant Status (TOS) analysis. This method is based on the conversion of ferrous ions to ferric form by oxidant compounds and the formation of complexes between the resulting ferric ions and xylenol orange. The resulting color change was measured spectrophotometrically, and the results were reported in µmol hydrogen peroxide equivalent/L [57]. The Oxidative Stress Index (OSI) was calculated using TAS and TOS values together; this parameter was expressed as a percentage (%) based on the TOS/TAS ratio. The OSI value was considered as an indicator that quantitatively reveals the balance between the antioxidant capacity and oxidant load of the extracts [58].

The free radical scavenging activities of the optimized mushroom extracts were determined using the commonly used DPPH (2,2-diphenyl-1-picrylhydrazyl) (Sigma-Aldrich, St. Louis, MO, USA) method. Before experimental analyses, each extract was dissolved in dimethyl sulfoxide to prepare stock solutions at a concentration of 1 mg/mL. Samples taken from the stock solutions were reacted with a specific volume of DPPH solution, and a mixture containing a low percentage of methanol was added to ensure the homogeneity of the reaction medium. Considering the light-sensitive nature of the DPPH radical, the prepared samples were incubated at room temperature under dark conditions. At the end of the incubation period, the color change resulting from the reduction of the DPPH radical by antioxidant compounds was measured at a wavelength of 517 nm, and the absorbance values were recorded. The obtained results were expressed by calculating them in mg Trolox equivalent/g extract (mg TE/g) [58].

The ferric ion reducing powers of the extracts were evaluated using the FRAP (Ferric Reducing Antioxidant Power) method. In this analysis, specific volumes of extract solutions were mixed with freshly prepared FRAP reagent. The FRAP reagent was prepared by combining three separate components in appropriate ratios: acetate buffer (pH 3.6), ferric chloride solution, and TPTZ. The extract-reagent mixtures were subjected to a short incubation at 37 °C to complete the reaction. During this process, the ability of the antioxidant compounds in the extracts to reduce ferric ions to the ferrous form was evaluated based on the resulting color intensity. Measurements were performed at a wavelength of 593 nm, and antioxidant capacity was calculated in mg Trolox equivalent/g extract (mg TE/g) using absorbance values [58].

3.6. Anticholinesterase Activity Tests

The acetylcholinesterase (AChE) and butyrylcholinesterase (BChE) inhibitory potentials of optimized mushroom extracts were evaluated using a modified protocol based on the Ellman method [59]. Galantamine hydrobromide (Sigma-Aldrich, St. Louis, MO, USA) was used as a positive control. To determine the concentration-dependent anticholinesterase effects, extract solutions were prepared in the range of 3.125–200 μg/mL. All analyses were performed in microplate format. Initially, 0.1 M phosphate buffer (pH 8.0) was added to each well, followed by the addition of AChE (Type VI-S, EC 3.1.1.7) or BChE (EC 3.1.1.8) (Sigma-Aldrich, St. Louis, MO, USA) enzyme solutions and the corresponding extract solutions. The reaction mixtures were pre-incubated at 25 °C for 10 min under light-free conditions to allow enzyme–extract interactions. Subsequently, 5,5′-dithiobis-(2-nitrobenzoic acid) (DTNB) and the appropriate thiocholine ester substrates, acetylthiocholine iodide or butyrylthiocholine iodide, were added to initiate the enzymatic reaction. The formation of the yellow-colored 5-thio-2-nitrobenzoate anion was monitored spectrophotometrically at 412 nm using a microplate reader. To eliminate potential interference caused by the intrinsic color of the extracts, appropriate extract blanks were included and absorbance values were corrected accordingly. Enzyme inhibition percentages were calculated from the corrected absorbance values, and anticholinesterase activities were expressed as half-maximal inhibitory concentration (IC₅₀) values (μg/mL).

3.7. Antiproliferative Activity Tests

The effects of the optimized mushroom extracts on cell proliferation were evaluated in vitro using three different human cancer cell lines: lung adenocarcinoma (A549), breast adenocarcinoma (MCF-7), and prostate carcinoma (DU-145), all obtained from the American Type Culture Collection (ATCC). To determine dose-dependent antiproliferative activity, extract solutions were prepared at concentrations of 25, 50, 100, and 200 µg/mL. Cells were cultured in appropriate growth media under standard culture conditions (37 °C, 5% CO₂) and detached using Trypsin–EDTA when they reached approximately 70–80% confluence. Cell suspensions were seeded into 96-well plates at a density of 1 × 10⁴ cells per well and pre-incubated for 24 h to allow attachment to the culture surface. All experimental conditions were performed in triplicate. Stock solutions of the optimized extracts were prepared in dimethyl sulfoxide (DMSO) and diluted with culture medium to obtain the desired final concentrations. The final DMSO concentration did not exceed 0.1% (v/v) in any well, and DMSO-treated cells were used as solvent controls. After treatment with the extracts for 24 h, the culture medium was removed, and cell viability was assessed using the MTT assay. MTT solution was added to each well at a final concentration of 0.5 mg/mL, and the plates were incubated at 37 °C for 4 h to allow the formation of purple formazan crystals in metabolically active cells. Following incubation, the formazan crystals were dissolved using DMSO, and absorbance values were measured at 570 nm using a microplate reader. Cell viability was expressed as a percentage relative to the control group, and the antiproliferative effects of the optimized fungal extracts on A549, MCF-7, and DU-145 cell lines were evaluated accordingly. Data are expressed as mean ± standard deviation (SD) of three independent experiments, and all measurements were performed in triplicate [58].

3.8. Phenolic Analysis

The phenolic composition of the optimized R. obtusissima extracts was analyzed using a Shimadzu LC–MS/MS-8030 system equipped with a binary pump (LC-20ADXR) and an autosampler (SIL-20ACXR) (Shimadzu Corporation, Kyoto, Japan). Chromatographic separation was achieved using a reverse-phase Inertsil ODS-4 C18 column (2.1 × 50 mm, 2 µm particle size), with the column oven (CTO-10ASvp) maintained at 40 °C throughout the analysis. The mobile phase consisted of solvent A (ultrapure water containing 0.1% formic acid) and solvent B (LC–MS grade methanol containing 0.1% formic acid), delivered in binary gradient mode at a total flow rate of 0.40 mL/min. The gradient program was set as follows: 5% B initially, increased to 95% B at 4.0 min, held at 95% B until 7.0 min, then returned to 5% B at 7.01 min and maintained until 12.0 min. The maximum system pressure was set to 660 bar. The injection volume was 2 µL. A total of 24 phenolic and related standard compounds representing different structural classes were simultaneously screened and quantified. The analyzed standards included acetohydroxamic acid, catechin hydrate, vanillic acid, syringic acid, thymoquinone, resveratrol, myricetin, kaempferol (positive ion mode), as well as fumaric acid, gallic acid, protocatechuic acid, 4-hydroxybenzoic acid, caffeic acid, salicylic acid, phloridzin dihydrate, 2-hydroxycinnamic acid, oleuropein, 2-hydroxy-1,4-naphthoquinone, naringenin, silymarin, quercetin, luteolin, alizarin, and curcumin (negative ion mode). Phenolic compounds were identified and quantified based on comparison with authentic reference standards using retention times and LC–MS/MS spectral data. Calibration curves were constructed for each standard compound, and limits of detection (LOD) and limits of quantification (LOQ) were calculated and are presented in Table 4. The applied chromatographic and mass spectrometric conditions ensured high selectivity, sensitivity, and reliable quantification of phenolic constituents in the optimized extracts.

3.9. Chemical Composition Analysis

The chemical component profiles of the optimized mushroom extracts were analyzed using gas chromatography–mass spectrometry (GC-MS). The analyses were performed in three independent replicates to identify volatile and semi-volatile compounds present in the extracts and determine their relative abundances. Chromatographic separation was performed using an HP-5 MS capillary column (30 m × 0.25 mm inner diameter, 0.25 µm film thickness). Helium was preferred as the carrier gas, and the flow rate was set to 3 mL/min. Sample injections were performed in split mode, and the injector temperature was kept constant at 280 °C. The oven temperature program was applied as follows: initially at 50 °C for 2 min, then the temperature was increased to 120 °C at a rate of 15 °C per minute and held at this temperature for 2 min. Subsequently, the temperature was increased to 300 °C at a rate of 5 °C per minute and kept constant at this value for 16 min. The mass spectrometer detector temperature was set to 280 °C. Compound identification was performed based on comparing the obtained mass spectra with commercial libraries and their agreement with literature data. The amounts of extract components were evaluated semi-quantitatively by considering the relative percentages of the peak areas. This approach allowed for the comparative determination of the chemical profiles of the extracts. The identification of lipophilic compounds detected by GC–MS was performed based on the comparison of mass spectral data with the NIST and Wiley spectral libraries, together with published literature data. As authentic reference standards were not employed, compound assignments should be considered tentative. Therefore, GC–MS results were primarily used to provide a comparative chemical profiling of the optimized extracts rather than definitive structural confirmation. This approach is commonly applied in exploratory phytochemical studies and serves to support biological activity interpretations.

3.10. Statistical Analysis

The statistical analysis of the experimental findings obtained in this study was performed using IBM SPSS Statistics version 21.0 software. Independent samples t-test was applied to evaluate the differences between the mean values of two independent groups. In analyses comparing three or more groups, one-way analysis of variance (One-Way ANOVA) method was used. In the evaluation of ANOVA results, the statistical significance level was accepted as α = 0.05 at a 95% confidence interval. If a significant difference was found between groups, Duncan’s multiple comparison test was applied to determine which groups were the source of these differences.

4. Conclusions

In this study, the chemical and biological properties of extracts obtained from R. obtusissima were comprehensively evaluated for the first time using artificial intelligence-assisted optimization approaches. The results clearly demonstrate that ANN-GA-based optimization offers a more effective strategy than the traditional RSM approach in enhancing antioxidant capacity as well as anticholinesterase and antiproliferative activities. The optimization strategy was deliberately designed to maximize antioxidant capacity as a model biological endpoint, rather than to exhaustively screen all possible classes of bioactive compounds. LC-MS/MS analyses revealed that ANN-GA optimization selectively enriched the phenolic profile, leading to significantly increased levels of biologically active compounds such as gallic acid, caffeic acid, and quercetin. In parallel, GC-MS data indicated that fatty acids and ester derivatives, which may support biological activity, were more prominently represented in the ANN-GA-optimized extract. Importantly, these findings indicate that the observed biological effects are not driven by single dominant molecules, but rather by optimization-induced shifts in chemical composition that promote synergistic interactions among multiple bioactive compound groups. Therefore, the biological relevance of R. obtusissima arises not from the novelty of individual compounds, but from its responsiveness to optimization-driven functional enrichment. Accordingly, the enhanced antioxidant capacity, reduced oxidative stress index, and stronger suppression of cancer cell proliferation observed in the ANN-GA extract can be directly attributed to the optimization-guided enrichment of functionally relevant phenolic and lipophilic constituents. From a broader perspective, R. obtusissima emerges as a promising natural resource with significant functional and biopharmaceutical potential, the expression of which can be substantially amplified through appropriate extraction optimization strategies. While isolation and definitive structural elucidation of individual phenolic compounds, along with assessment of their standalone biological activities, represent necessary next steps toward pharmaceutical product development, the present study establishes a robust scientific foundation by demonstrating how AI-assisted optimization can effectively tailor chemical profiles and maximize biological efficacy. In this context, the findings provide a solid methodological and exploratory basis for future compound-level investigations, functional food applications, and the rational design of biologically active natural extracts.

Abbreviations

The following abbreviations are used in this manuscript:

DMSO	Dimethyl sulfoxide
DTNB	5.5″-dithiobis-(2-nitrobenzoic acid)
MTT	(3-[4,5-dimethylthiazol-2-yl]-2,5-diphenyl-tetrazolium bromide)
OSI	Oxidative stress index
TAS	Total Antioxidant Status
TOS	Total Oxidant Status

Author Contributions

Conceptualization, İ.K., M.S. and I.A.; methodology, İ.K., M.S. and I.A.; software, İ.K., M.S. and I.A.; validation, İ.K., M.S. and I.A.; formal analysis, İ.K., M.S. and I.A.; investigation, İ.K., M.S. and I.A.; resources, İ.K., M.S. and I.A.; data curation, İ.K., M.S. and I.A.; writing—original draft preparation, İ.K., M.S. and I.A.; writing—review and editing, İ.K., M.S. and I.A.; visualization, İ.K., M.S. and I.A.; supervision, İ.K., M.S. and I.A.; project administration, İ.K., M.S. and I.A. All authors have read and agreed to the published version of the manuscript.

Institutional Review Board Statement

Not applicable.

Informed Consent Statement

Not applicable.

Data Availability Statement

If requested, data related to the study may be requested from the corresponding author upon rea-sonable request.

Conflicts of Interest

The authors declare no conflicts of interest.

Funding Statement

This research received no external funding.